Bay HD comments

[Ranger]

It's rich that you would chastize me for using a 2006 interview about TL HD DVD discs (which is pertinent because both HD45 and HD51 are both triple layer discs and the interview was with a TOSHIBA engineer), then you would turn around and suggest that MPEG-2 is still the most commonly used codec on BLu-ray. Now there isn't anything wrong with MPEG-2, but to suggest that it is STILL the most commonly used codec is typical HD DVD Propagana Group type stuff.

Let's look at 2007 shall we, the last 48 or so weeks.

MPEG-2 65 23.30%
AVC 132 47.31%
VC-1 82 29.39%

Still the most commonly used? I don't think so. Perhaps there are still more discs out there with MPEG-2, but only a fool would suggest that's not changing rapidly with all the great AVC encodes Panasonic are churning out for Disney and Fox and Sony's great AVC ecnodes as well.

Doby, you know better. I posted recently that Sony's Mpeg2 encoding equipment investment was a big reason for making sure Blu-ray had extra storage and bandwidth. It was needed to allow Mpeg2 to deliver the same sort of image quality that can be achieved with AVC/Mpeg4 and VC1.

As further proof points on this we saw Sony Pictures release all of their titles in Mpeg2 in 2006 and Don Eklund telling us that Mpeg2 was better than AVC/Mpeg4 and VC1. Now that Sony is shipping new AVC/Mpeg4 encoding equipment, they no longer need to sell us on the superiority of Mpeg2......they are now mostly releasing in AVC/Mpeg4.

Funny coincidence huh? When they only have Mpeg2 encoding equipment to sell, it's the best codec there is. Now that they have AVC/Mpeg4 encoding equipment for sale......well you get the picture.

I've posted this all before, but I'm going to assume you didn't catch it as this thread moves fast.

[The Drizzle]

Doby, you know better. I posted recently that Sony's Mpeg2 encoding equipment investment was a big reason for making sure Blu-ray had extra storage and bandwidth. It was needed to allow Mpeg2 to deliver the same sort of image quality that can be achieved with AVC/Mpeg4 and VC1.

As further proof points on this we saw Sony Pictures release all of their titles in Mpeg2 in 2006 and Don Eklund telling us that Mpeg2 was better than AVC/Mpeg4 and VC1. Now that Sony is shipping new AVC/Mpeg4 encoding equipment, they no longer need to sell us on the superiority of Mpeg2......they are now mostly releasing in AVC/Mpeg4.

Funny coincidence huh? When they only have Mpeg2 encoding equipment to sell, it's the best codec there is. Now that they have AVC/Mpeg4 encoding equipment for sale......well you get the picture.

I've posted this all before, but I'm going to assume you didn't catch it as this thread moves fast.

Sounds like business to me. VC1 and AVC are damn similar and have their good and bad points.

[Ranger]

I have read this very long thread in the hopes that Mr. Bay would elaborate as to why he feels that Blu Ray suits his movies better than HD DVD. I certainly respect his views, but was curious to know what technical reason
led him to this stance. Unfortunately, I just saw people posting their guesses on why, such as "more color space", "higher bandwidth", "higher bit rate", "more capacity".

While I hope Mr. Bay will revisit this thread and offer more background to his
preference for Blu Ray, I would like to address the potential technical advantages that Blu Ray supporters tout in regards to an alleged difference in picture quality over HD DVD.

Color Space--to my knowledge, ALL Blu Ray and HD DVD discs have been authored with 4:2:0 color space, regardless of codec or bit rate. Same as
all DVDs.

Higher Bit Rate--yes, Blu Ray has the potential for about 10Mbps more of raw video stream than HD DVD. This is important when encoding with MPEG2.
However, VC-1 and H.264(AVC) modern codecs are optimized for low bit
rates. They were designed to use HALF the bit rate of MPEG2 and yield
equal or better image quality. So a Blu Ray title encoded in MPEG2 at
the maximum video transfer rate of 40Mbps would be considered ample
for that codec. AVC and VC-1 only require 20Mbps for equal quality,
much lower than HD DVD's maximum video transfer rate of 30Mbps.

So, we know that color space and transfer rate using low bit rate optimized
codecs does not favor Blu Ray over HD DVD in image quality. Indeed, I haven't heard of one Blu Ray title that looks better than the same title on
HD DVD. Some would say that is because the title was encoded at a lower
bit rate for HD DVD's limits and used for the Blu Ray title, but I maintain that throwing more bit rate at a modern codec like VC-1 or AVC would not make a difference visually. If there are already no visible compression artifacts at 20 or 30Mbps, encoding a disc at 40Mbps won't yield any discernable quality advantage.

I continue to be curious about what it is that Mr. Bay sees in regards to an image quality difference between Blu Ray and HD DVD given the same title,
same transfer and same codec.

Dino

Very well put. I share your curiosity.

[Mukha]

The last few pages appear to have been spent debating a deterministic mathematical algorithm (and a very simple one at that).

IVTC is the perfect reconstruction of an original 1080p24 source from a 1080i60 signal that has undergone 3:2 pulldown. The Telecine/Inverse Telecine process is performed without loss. As it relates to HDM, the Toshiba A3 HD DVD player can only output content at 1080i60 therefore for the vast majority of content that is encoded on HD DVD in 1080p24 (film) it has to perform 3:2 pulldown to get 1080i60. If this signal is fed to a display that is capable of 1080p24 playback (it might frame double 1080p24 to 1080p48) then it will perform IVTC on this 1080i60 signal to reconstruct without loss the 1080p24 original. The fact that it is a perfect reconstruction of the original 1080p24 source is not up for debate. There is no temporal smearing of fields as there is in a interlaced signal as the 1080i60 signal contains both fields from each original frame allowing reconstruction of the original frame.

[wingzero]

I believe that every HD player presently interlaces the images before either sending it to the display or reinterlacing it. The lower priced HD DVD players don't have the deinterlacer that all the other HD players have. And, i and p are really the same thing. All that is different is the way the information is displayed and how fast it refreshes. Interlaced video refreshes twice as fast, obviously. Oh and your question is ridiculous. In 99% of cases the interlacing is seemless no matter where it occurs. The video mixers will most likely prevent pure video from passing over the buss, though it would be cool if it could be done and we could bypass the features when we cant, something like they presently do with the audio in 1.3 players (except the PS3 which lacks the silicon).

?? Are you serious ? For you interlaced and progressive are the same thing ? Even cheap $30 chinese DVD players do have motion adaptive deinterlacing of some sort.

[wingzero]

No, I am trying to keep misinformation from being spread. He's been calling me childish over and over even though he has been posting incorrect information, and now you go the "childish" direction too.

Honestly, I think it would be nice if wingzero learned something here. Especially since he still doesn't seem to comprehend the main reason for 1080p24sf. But with his attitude, I'm doubting he will learn anything and will keep stating that anybody who thinks a progressive source can be interlaced and then deinterlaced without loss is wrong. He seems to have learned what motion adaptive deinterlacing is, but has misapplied it, thinking that it is needed for something where it is not.

--Darin

I gave you a link to a Philips technical document that gives a pretty detailed overview of interlacing and how modern algorithms work to solve its issues. You kept repeating incorrect information from websites and forums that really spread wrong technical info for the general public.
You lack the basic knowledge of how things work but you are too arrogant to admit it. That's the reason why you are acting so childish.
Accusing me of spreading misinformation is pathetic, then you are telling that all developers books and the whole literature in the field is a fake because you believe that telecine pulldown is magic and can solve everything better. This is acting childish.

[wingzero]

I can understand that, but as I understand it all players are currently being forced to interlace at some point. I expect that whatever format(s) we are finally dealing with, the present players and discs will seem inferior to the players and discs we get in the next couple of years. Look at DVDs development. The first players and discs were horrible.

You are wrong. Even cheap $30 DVD players with a Mediatek chipset/DSP (or a Zoran or other brand) can output full progressive MPEG-2 video stream if you set the option in the menu.

[wingzero]

No, I am trying to keep misinformation from being spread. He's been calling me childish over and over even though he has been posting incorrect information, and now you go the "childish" direction too.

Honestly, I think it would be nice if wingzero learned something here. Especially since he still doesn't seem to comprehend the main reason for 1080p24sf. But with his attitude, I'm doubting he will learn anything and will keep stating that anybody who thinks a progressive source can be interlaced and then deinterlaced without loss is wrong. He seems to have learned what motion adaptive deinterlacing is, but has misapplied it, thinking that it is needed for something where it is not.

--Darin

Ask Nelson or Mr.Bay to explain interlacing to you or have an employee, a codec designer, an engineer working for ILM maybe come here and explain things to you if they have a bit of time to waste. I'd be amazed if any of them would ever tell that an interlaced video could be reconstructed to the original progressive source with no approximation error and no information loss by using simple outdated telecine pulldown, really...

[wingzero]

The last few pages appear to have been spent debating a deterministic mathematical algorithm (and a very simple one at that).

IVTC is the perfect reconstruction of an original 1080p24 source from a 1080i60 signal that has undergone 3:2 pulldown. The Telecine/Inverse Telecine process is performed without loss. As it relates to HDM, the Toshiba A3 HD DVD player can only output content at 1080i60 therefore for the vast majority of content that is encoded on HD DVD in 1080p24 (film) it has to perform 3:2 pulldown to get 1080i60. If this signal is fed to a display that is capable of 1080p24 playback (it might frame double 1080p24 to 1080p48) then it will perform IVTC on this 1080i60 signal to reconstruct without loss the 1080p24 original. The fact that it is a perfect reconstruction of the original 1080p24 source is not up for debate. There is no temporal smearing of fields as there is in a interlaced signal as the 1080i60 signal contains both fields from each original frame allowing reconstruction of the original frame.

Telecine/Inverse Telecine can't reconstruct anything without loss. It works by discarding frame/fields and doubling some in the process based on the source and target frame rates.
And you are wrong, you too don't have a clue of what it's going on to the signal in the frequency domain and how the optical flow axis gets corrupted due to interlacing.
What is not up for debate is that telecine pulldown is an outdated technique but everyone here believes that what some websites and forums of marketeers keep telling you is the truth. The only truth is that you can find in developers books and official multinationals documents working in the field and spending billions on R&D. Nowadays practically every DVD player and more expensive hardware is using some sort of motion compensation techniques at least. Thinking otherwise it's just wrong.

[darinp2]

wingzero,

How about answering my previous post that went:

Back to one of your claims:

2) Segmenting a frame into fields is interlacing when the split image is sent even(odd) fields first and odd(even) after those. Deinterlacing with no loss of quality is not possible.

I ask again:

If I sent you a picture by sending the odd lines today and the even lines tomorrow, could you reconstruct the original?

Please tell me that you would be smart enough to be able to reconstruct the original if I did that. And that if sent you the odd lines today, the same odd lines tomorrow, and then the even lines the day after tomorrow (telling you that this is what I would be doing), that you could discard one of the odd fields and still reconstruct the original. If you think you couldn't figure out how to reconstruct the original under those conditions, please explain why you think you couldn't.
---
It is quite clear that you don't understand this stuff when you claim that a segmented frame cannot be reinterlaced with no loss in quality. Maybe you should ask ILM or somebody else you know what 1080p24sf is and why it was created, since you don't seem to comprehend it.

Your posts seem a little bit like a person just learning calculus and thinking that it must be used if somebody wants to balance their checkbook, since addition and subtraction are outdated. Use the right tool for the job. Motion compensation is not for progressive sources that are just sent in interlaced signals with enough bandwidth for the original information.

You've already been wrong when you used the bandwidth angle to try to claim that there wouldn't be enough bandwidth. If you knew as much about this subject as you seem to think you do, you wouldn't have made that mistake.

You can link to all the articles you want, but if you don't understand what they relate to, then it doesn't do you any good. That article was for interlaced signals where information for the full frames is missing, so advanced techniques need to be used to try to figure out what the missing information might have been.

Now, how about it. If I sent you the odd lines for a picture today and the even lines tomorrow, could you figure out how to reconstruct the original picture? Is there a reason you are avoiding answering that?

Telecine/Inverse Telecine can't reconstruct anything without loss. It works by discarding frame/fields and doubling some in the process based on the source and target frame rates.

With 1080p24 source and 1080i60 signals it works by repeating some information and then discarding the repeated information. You seem to be thrown off by that, as if you think that would mean something must be missing, when it is just the repeated information that gets thrown away when this is done correctly. Maybe that is why you incorrectly thought that the bandwidth was going to be a problem for 1080p24 sources and 1080i60 signals.

You sure seem to want to avoid answering simple and straightforward questions like:

Let's go back to a basic here though. If no filtering was added for the 1080p24 to 1080i60 stage in the player, would you still argue that the original 1080p24 could not be reconstructed?

Why is that?

--Darin

[Mukha]

The fact that it is a perfect reconstruction of the original 1080p24 source is not up for debate.

Telecine/Inverse Telecine can't reconstruct anything without loss. It works by discarding frame/fields and doubling some in the process based on the source and target frame rates.

I get the feeling that you seem to think 3:2 pulldown and it's inverse are far more complicated than they actually are. A 1080p24 source that has undergone 3:2 pulldown is still fundamentally progressive it is just being represented in an interlaced form and IVTC has nothing to do with common deinterlacing techniques. It can be mathematically proven that IVTC back to a 1080p24 source is lossless. No fields are discarded except those that are duplicates. It's nothing more than addition and subtraction. There can be no debate about this.

I think you are either trolling or you are way out of your depth and this side issue debating something that fundamentally should not be debated has completely derailed this thread. Darinp2 has respectfully tried to provide the facts about the matter even in the face of continual insults and I've learnt a lot here from him (thanks darin!) but this whole topic should be moved to another thread as this really doesn't particularly have much do to with any differences between the HDM formats.

[The Drizzle]

Do you guys understand that there are tens of millions of displays that will not accept 1080p and/or have no better than component (not that component is bad) and can not have new cabling pulled to them. I personally have hundreds of installs like that. HDMI wasn't an option for 12 of the 15 years I've been installing/designing. Of course if you have a display that will take 1080p that is the best option in most cases, but if not an A1, A2, or A3 is just as good as a PS3 or the Pioneers. JUst a little perspective.

[wingzero]

I get the feeling that you seem to think 3:2 pulldown and it's inverse are far more complicated than they actually are. A 1080p24 source that has undergone 3:2 pulldown is still fundamentally progressive it is just being represented in an interlaced form and IVTC has nothing to do with common deinterlacing techniques. It can be mathematically proven that IVTC back to a 1080p24 source is lossless. No fields are discarded except those that are duplicates. It's nothing more than addition and subtraction. There can be no debate about this.

So you join Darinp2 in spreading incorrect information. If you apply 3:2 pulldown you destroy the original progressive source, it's not progressive anymore. Discarding any field or frame causes distortions to the video signal and the optical flow axis gets corrupted. There is no "progressive in interlaced form" thing. It's like telling that the Earth is flat, really.

It's absolutely crazy what you are claiming here, really. Along with all the two different interlacing definitions Darinp2 even came up with tying to prove his wrong points.

I think you are either trolling or you are way out of your depth and this side issue debating something that fundamentally should not be debated has completely derailed this thread. Darinp2 has respectfully tried to provide the facts about the matter even in the face of continual insults and I've learnt a lot here from him (thanks darin!) but this whole topic should be moved to another thread as this really doesn't particularly have much do to with any differences between the HDM formats.

If you don't understand the basics of video coding and you don't understand how interlacing and telecine pulldown work, then you surely can't debate about the other formats specifications nor understand why Blu-Ray is better than HD-DVD. You and Darinp2 and others here completely lack the basic theory to understand what is going on to a video signal and how it gets processed at various stages.

Accusing me of trolling and other childish accusations it's absolutely pathetic. Darinp2 more than you just doesn't allow others to know any technical argument better than he does, he is acting like a little kid. Insulting me will not prove either you or him right. Both of you should study some basic developers book on video coding and signals theory before going around accusing others of being ignorant. You are the ones that not only don't understand how things work but you don't accept the fact that every developer book would tell you the exact same things I kept telling you here and that you don't want to listen to.

Also, the fact that you just joined the forum to start attacking me on the subject and claiming the same exact wrong information that Darinp2 is repeating over and over and telling that he is right.. it makes me wonder if you are just another account of the same guy behind the Darinp2 nickname... Well, given his childish behaviour so far I wouldn't be that much surprised if that was the case.


Although it's probably just another waste of time (because you don't want to study things and you don't care, you just want me and everyone else to say that you/Darinp2 are/is right while the truth is that you/him are wrong), some more links to provide some examples showing how many resources are spent on R&D to solve interlacing issues and provide better de-interlacing techniques to minimize the approximation errors trying to reconstruct an original progressive signal. All of them involve the use of some sort of motion detection technique along with the needed spatio-temporal interpolation to perform the task. At least I hope that the links providing correct information on interlacing and how it works that I paste here will be useful to anyone willing to listen and study a bit on the subject instead of trying to insult me and acting like a little whining kid.



--
http://search.ieice.org/bin/summary....7&lang=E&abst=


Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences Vol.E90-A No.11 pp.2575-2583
Publication Date: 2007/11/01
Online ISSN: 1745-1337
Print ISSN: 0916-8508
Type of Manuscript: PAPER
Category: Image

Summary:
Scene changes occur frequently in film broadcasting, and tend to destabilize the performance with blurred, jagged, and artifacts effects when de-interlacing methods are utilized. This paper presents an efficient VLSI architecture of video de-interlacing with considering scene change to improve the quality of video results. This de-interlacing architecture contains three main parts. The first is scene change detection, which is designed based on examining the absolute pixel difference value of two adjacent even or odd fields. The second is background index mechanism for classifying motion and non-motion pixels of input field. The third component, spatial-temporal edge-based median filter, is used to deal with the interpolation for those motion pixels. Comparing with the existed de-interlacing approaches, our architecture design can significantly ameliorate the PSNRs of the video sequences with various scene changes; for other situations, it also maintains better performances. The proposed architecture has been implemented as a VLSI chip based on UMC 0.18-µm CMOS technology process. The total gate count is 30114 and its layout area is about 710 710-µm. The power consumption is 39.78 mW at working frequency 128.2 MHz, which is able to process de-interlacing for HDTV in real-time.

-----


http://ieeexplore.ieee.org/Xplore/lo...number=1597157

High performance de-interlacing algorithm for digital television displays
Gwo-Long Li; Mei-Juan Chen
Display Technology, Journal of
Volume 2, Issue 1, March 2006 Page(s): 85 - 90
Digital Object Identifier 10.1109/JDT.2005.864157
Summary: In this paper, a high performance de-interlacing algorithm is proposed to reduce the artifact effects produced by interlaced scanning TV systems. In the proposed algorithm, a moving-stationary detector is presented to determine where the missing pixels belong to moving or stationary region. If the missing pixel belongs to the stationary region, the temporal-wise interpolation is adopted; otherwise, the spatial-temporal-wise interpolation is applied. In addition, the proposed algorithm has simple operations chiefly involves addition and subtraction. This simple computation conducts the proposed algorithm to real-time applications efficiently. Experimental results show that the proposed algorithm can improve the subjective and objective qualities compared with other de-interlacing algorithms.

-----

[Ranger]

Do you guys understand that there are tens of millions of displays that will not accept 1080p and/or have no better than component (not that component is bad) and can not have new cabling pulled to them. I personally have hundreds of installs like that. HDMI wasn't an option for 12 of the 15 years I've been installing/designing. Of course if you have a display that will take 1080p that is the best option in most cases, but if not an A1, A2, or A3 is just as good as a PS3 or the Pioneers. JUst a little perspective.

Apparently not.

[darinp2]

wingzero,

Are you going to respond to my last post, or avoid the questions in there to try to keep others from seeing how wrong you are? It is amazing how many times you have claimed others are childish while you won't even answer simple questions like:

"If I sent you a picture by sending the odd lines today and the even lines tomorrow, could you reconstruct the original?"

and

"Let's go back to a basic here though. If no filtering was added for the 1080p24 to 1080i60 stage in the player, would you still argue that the original 1080p24 could not be reconstructed?"

A child might have problems understanding that duplicate data can be discarded without losing anything, but after you incorrectly used the lack of enough bandwidth argument, you have now moved to claiming that discarding any field or frame messes things up, when the fact is that no field or frame itself is discarded in the case we are discussing (1080p24 sent with 1080i signals), only duplicates are discarded. Do you seriously have trouble with the concept that throwing away duplicate information doesn't cause a problem, or are you just trying to find anything to support your original claim?

Hopefully someday you will learn enough about this subject to not make mistakes like your claim:

2) Segmenting a frame into fields is interlacing when the split image is sent even(odd) fields first and odd(even) after those. Deinterlacing with no loss of quality is not possible.

If you get a chance to talk to an expert, ask them this question:

"If a 1080p24 source is sent with a 1080p24sf signal instead of a 1080p24 signal, can the original progressive information be reconstructed."

You obviously said that it can't, but you are wrong and it is just one piece of evidence that you don't really understand the subject that well. You understand it well enough to know that very advanced techniques are needed in some cases, but not enough to understand when they aren't needed. Basically goes along with your childish comments, except in the opposite direction, with somebody who doesn't understand when advanced stuff isn't needed and refuses to answer simple questions because it might make it real obvious that they don't understand the subject that well, is the one being childish.

As I said before:

You can link to all the articles you want, but if you don't understand what they relate to, then it doesn't do you any good.

Companies spend a lot of money on things that involve things like imaginary numbers and Fourier Transforms, but it doesn't mean that they are needed for balancing a checkbook. Just like here, motion compensation isn't needed for 1080p24 packaged in a 1080p24sf or 1080i60 signal type (unless the goal is to change from the original, like taking that 24p after reconstruction and converting to 48p or 72p).

Although this is getting really old, I'm guessing others can see that there is a reason you won't answer:

If I sent you a picture by sending the odd lines today and the even lines tomorrow, could you reconstruct the original?

And please don't stoop to answering that and then claiming that throwing away the duplicate data is the problem here, after I already asked if you could reconstruct the picture if I sent you the odd lines twice and the even lines once.

--Darin

[wingzero]

Darinp2, listen, it's getting really boring to explain things to you over and over and you keep repeating the exact same wrong statements without understanding anything I told you.
After having came up with the insane idea of the absurd existence of two different types of interlacing that you claimed to exist along with many other silly statements of yours on the subject, now you are even telling that Fourier Transform is not needed ... That is such a silly statement that it's really crazy. If Fourier transform, Fast Fourier Transform and Discrete Cosine Transform didn't exist then nowadays there would be no digital video coding and it's highly unlikely that even any analogue video standard could have been developed in the first place.
The fact that you don't know how this stuff works it doesn't mean that developers working in the field are wasting time,money and resources while you have the absolute truth. For you and others it seems that telecine pulldown solved everything and interlacing it's not an issue that needs better solutions.
Well, you are wrong, face it.

You can keep telling all the incorrect info you want but you are wrong. Your claims are wrong and it's not the professional researchers and developers, video architects, codec designers and video engineers working in the field the ones being wrong, it's you and others ideas both here and on some well known websites and forums that are spreading wrong information and can't accept the truth that's written on developers book and official documents like those published by ISO/IEC and IEEE organizations.

As I suspected, it was a waste of time trying to let you understand anything provided by official documents of proposed and implemented de-interlacing algorithms. You don't accept the simple fact that you need to study a lot before trying to debate anything on the subject. Counting frame/fields beind discarded and thinking that telecine pulldown is what is all about to understand interlacing and video coding it's absolutely wrong. Also, your way of answering telling that I would be the ignorant one it's the most pathetic and childish thing you could have done. The few technical documents links I provided are already enough to let anyone see how things really work and what interlacing means, but there are tons of developers books telling the exact same things that anyone could buy and read.

[darinp2]

Darinp2, listen, it's getting really boring to explain things to you over and over and you keep repeating the exact same wrong statements without understanding anything I told you.

I understand what you told me, but also understand enough to know that it doesn't apply to the situation being discussed, while you don't understand enough to realize that. And your last post pretty much confirmed that you will avoid simple question where the answer would lead to people seeing that you are wrong:

Are you going to respond to my last post, or avoid the questions in there to try to keep others from seeing how wrong you are? It is amazing how many times you have claimed others are childish while you won't even answer simple questions like:

"If I sent you a picture by sending the odd lines today and the even lines tomorrow, could you reconstruct the original?"

It is getting old because you don't have enough integrity to answer those simple questions and are trying to find any angle you can to claim that 1080p24 can't be sent over a 1080i60 signal and then get reconstructed correctly. You've already failed with the not enough bandwidth, the codec does filtering, and fields/frames are getting thrown away arguments, since there is enough bandwidth, any codec filtering applies to both 1080p24 and 1080i60 output signals, and the stuff getting throw away is just duplicate information. Your argument that experts work on deinterlacing algorithms proves that inverse telecine doesn't work is also bogus, because the reason those experts work on that stuff is because there are many cases where it can't be used (like video shot with 1080i) and what they do is very useful for the situations it applies to (not the situation being discussed here). Funny that you complain about this getting old when you still don't even understand 1080p24sf and have left your false claim out there about it. Learn what it is and what can be done with it, and then you will be on the road to understanding this subject matter.

After having came up with the insane idea of the absurd existence of two different types of interlacing that you claimed to exist along with many other silly statements of yours on the subject, now you are even telling that Fourier Transform is not needed ... That is such a silly statement that it's really crazy.

Did you really have trouble comprehending what I said:

Companies spend a lot of money on things that involve things like imaginary numbers and Fourier Transforms, but it doesn't mean that they are needed for balancing a checkbook.

Seemed so simple a child could understand it to me, but somehow you thought that sentence said that imaginary numbers and Fourier Transforms aren't needed? You don't honestly believe they are needed for balancing a checkbook do you? Seems simple to understand that they are needed in some cases, but not in others.

The fact that you don't know how this stuff works it doesn't mean that developers working in the field are wasting time,money and resources while you have the absolute truth.

They are not wasting their time because there are situations where what they are doing is very useful (which is why they do it). The smart ones know not to use that stuff in cases where it doesn't apply though.

For you and others it seems that telecine pulldown solved everything and interlacing it's not an issue that needs better solutions.

Nope. There are cases like the one being discussed here where that other stuff isn't needed, but many cases where it is.

You can keep telling all the incorrect info you want but you are wrong. Your claims are wrong and it's not the professional researchers and developers, video architects, codec designers and video engineers working in the field the ones being wrong ...

No, those people aren't wrong. You are. They understand that there are different situations (at least the smart ones do), but you don't.

As I suspected, it was a waste of time trying to let you understand anything provided by official documents of proposed and implemented de-interlacing algorithms.

Let's go back to your last couple, because it shows that you don't understand when those things apply and when they don't.

http://search.ieice.org/bin/summary....7&lang=E&abst=


Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences Vol.E90-A No.11 pp.2575-2583
Publication Date: 2007/11/01
Online ISSN: 1745-1337
Print ISSN: 0916-8508
Type of Manuscript: PAPER
Category: Image

Summary:
Scene changes occur frequently in film broadcasting, and tend to destabilize the performance with blurred, jagged, and artifacts effects when de-interlacing methods are utilized. This paper presents an efficient VLSI architecture of video de-interlacing with considering scene change to improve the quality of video results. This de-interlacing architecture contains three main parts. The first is scene change detection, which is designed based on examining the absolute pixel difference value of two adjacent even or odd fields. The second is background index mechanism for classifying motion and non-motion pixels of input field. The third component, spatial-temporal edge-based median filter, is used to deal with the interpolation for those motion pixels. Comparing with the existed de-interlacing approaches, our architecture design can significantly ameliorate the PSNRs of the video sequences with various scene changes; for other situations, it also maintains better performances. The proposed architecture has been implemented as a VLSI chip based on UMC 0.18-µm CMOS technology process. The total gate count is 30114 and its layout area is about 710 710-µm. The power consumption is 39.78 mW at working frequency 128.2 MHz, which is able to process de-interlacing for HDTV in real-time.

-----


http://ieeexplore.ieee.org/Xplore/lo...number=1597157

High performance de-interlacing algorithm for digital television displays
Gwo-Long Li; Mei-Juan Chen
Display Technology, Journal of
Volume 2, Issue 1, March 2006 Page(s): 85 - 90
Digital Object Identifier 10.1109/JDT.2005.864157
Summary: In this paper, a high performance de-interlacing algorithm is proposed to reduce the artifact effects produced by interlaced scanning TV systems. In the proposed algorithm, a moving-stationary detector is presented to determine where the missing pixels belong to moving or stationary region. If the missing pixel belongs to the stationary region, the temporal-wise interpolation is adopted; otherwise, the spatial-temporal-wise interpolation is applied. In addition, the proposed algorithm has simple operations chiefly involves addition and subtraction. This simple computation conducts the proposed algorithm to real-time applications efficiently. Experimental results show that the proposed algorithm can improve the subjective and objective qualities compared with other de-interlacing algorithms.

-----

I've highlighted the appropriate parts. The first one is for cases where the adjacent odd and even fields differ in time. Not the case being discussed here where each odd field has a corresponding even field that represents the same point in time, and vis versa (and you would understand that about 1080p24sf if you understood what that is). The 2nd one is for "interlaced scanning TV systems." Again, not what we are discussing here. What we are discussing here is 1080p24 with a 1080i60 signal type (or 1080p24sf) to a progressive display. Interlaced displays have the same problem with 1080p24 signal types (since they can't display it progressively) and what we are discussing here is the difference between sending 1080p24 with 1080i60 or 1080p24 signal types.

--Darin

[wingzero]

Darinp2 you are so arrogant and childish that you really don't know when it's time to say that you were wrong and apologize for insulting and accusing others of being ignorant while you are the one claiming incorrect things and mixing things up like a vortex to build up confusion in a hope to prove you right.
That's a childish attitude and it's getting silly wasting time repeating the same exact things to you. Is that a tactic of yours, perhaps ? Do you really believe that by repeating the same incorrect things over and over will make you look smart and cool and professional although the whole literature in the field simply proves you wrong ?

Now you keep repeating 1080p24sf as it was the magic word "to win against me" like this was a game to show how smart and cool you are. That's a childish attitude, indeed. 1080p24sf it's a marketing way of describing just another technique of multipling the video scan rate to reduce flickering, it has nothing to do with interlacing per se. Films in the theatre are progressive sources and techniques like the early double scan rate one weren't interlacing. Either doubling the scan rate working on fields or frames did and does work, however motion compensation and better techniques have been applied to scan rate just like it happened to de-interlacing. Simply because that does provide a better overall result than just rendering the same field or full frame twice on screen. Other studies that involve looking at the signals in the frequency domain and mapping it to an n-Dimensions space have been made to prove that and go beyond the simple double,triple (or higher) scan rate techniques.

Claiming that field/frames being discarded by telecine pulldown would be redundancy it's a naive statement. It really shows that you don't know what you are talking about. You are the one that needs to study on the subject a lot, and not me. But you are so arrogant and childish that you will never accept that and you will keep going around telling everyone that you are an expert in the field although you fail to understand the basics. Your attitude has been so rude for the whole thread, I must be really out of mind for having wasted so much time repeating over and over the same things to you. And you keep insulting and whining and acting childish like we were in a kindergarten and you were a little kid crying out loud because someone told you that you were wrong and politely tried to explain it to you. Geez, grow up !

[darinp2]

Darinp2 you are so arrogant and childish that you really don't know when it's time to say that you were wrong and apologize for insulting and accusing others of being ignorant while you are the one claiming incorrect things and mixing things up like a vortex to build up confusion in a hope to prove you right.
That's a childish attitude and it's getting silly wasting time repeating the same exact things to you. Is that a tactic of yours, perhaps ? Do you really believe that by repeating the same incorrect things over and over will make you look smart and cool and professional although the whole literature in the field simply proves you wrong ?

Now you keep repeating 1080p24sf as it was the magic word "to win against me" like this was a game to show how smart and cool you are. That's a childish attitude, indeed. 1080p24sf it's a marketing way of describing just another technique of multipling the video scan rate to reduce flickering, it has nothing to do with interlacing per se. Films in the theatre are progressive sources and techniques like the early double scan rate one weren't interlacing. Either doubling the scan rate working on fields or frames did and does work, however motion compensation and better techniques have been applied to scan rate just like it happened to de-interlacing. Simply because that does provide a better overall result than just rendering the same field or full frame twice on screen. Other studies that involve looking at the signals in the frequency domain and mapping it to an n-Dimensions space have been made to prove that and go beyond the simple double,triple (or higher) scan rate techniques.

Claiming that field/frames being discarded by telecine pulldown would be redundancy it's a naive statement. It really shows that you don't know what you are talking about. You are the one that needs to study on the subject a lot, and not me. But you are so arrogant and childish that you will never accept that and you will keep going around telling everyone that you are an expert in the field although you fail to understand the basics. Your attitude has been so rude for the whole thread, I must be really out of mind for having wasted so much time repeating over and over the same things to you. And you keep insulting and whining and acting childish like we were in a kindergarten and you were a little kid crying out loud because someone told you that you were wrong and politely tried to explain it to you. Geez, grow up !

All from somebody acting so childish that they still won't answer:

If I sent you a picture by sending the odd lines today and the even lines tomorrow, could you reconstruct the original?

You are becoming quite laughable. Please, show some integrity and answer that question, or quit complaining about wasting your time, since you are the one extending this conversation by making false claims and avoiding any simple question that would lead to you realizing how wrong you are.

--Darin

[xAgonyxScenex]

What do you think, Microsoft is going to give you the slightly compressed 1080p file?

Please. I downloaded Superman Returns on XBox live and it was all but 2GB that downloaded in 30min at a 720p resolution.




See above.



iTunes proves that people don't mind having the digital versions of albums, movies, and books.

It's not a matter of when, but how someone perfects it... like Apple did with music.

You don't think a very large part of the population is going to want to full 1080p HD files. Companies should absolutely be giving consumers a 1080p file for a movie download. And on top of that movies through XBOX live are 7 gigs not 2. Right now there is just not a viable way to download full high def movies in a way that is easy to become portable. And I'm sorry, but internet speeds are not fast enough right now in an affordable way to quickly download 7gigs for a 720p HD movie or a 15gig file for a 1080p. I don't believe for a second you got 7 gigs to download in 30mins.

A lot of a DVD costs comes from production of the actual discs and boxes. When you cut that out movies become cheaper. People will then buy more. On top of that movies always become cheaper anyways. I currently own 200DVD's. How would I ever store what would be basically 3 terabytes of space in a way that I could easily take to a friends place for movie night. How many people go to Wal-Mart and stock up on 2 for $10 movies ever weekend. The amount of space to store that kind of downloading is enormous.

My point is, is that it's going to take a long while before the method is perfected to the point of it over taking actual discs. Microsoft no doubt has a plan for when digital downlaod take over. That is absolutely where the future is headed. It's certainly not going to be in the next couple of years, which is what Michael seems to be getting at.

I realize it wont happen until someone perfects it like Apple did with iTunes. Thats a given man.

[wingzero]

Darinp2, I tried to explain what interlacing really is about to you just too many times here.
Your childish way of insulting me it's not making you look cool and smart, trust me. However I bet that you won't stop because you keep being so arrogant that you can't cut it out and apologize politely in any case, so it seems. Instead you keep attacking and insulting me of doing what you really did and accusing me of the wrong things you keep claiming, in a hope that those reading the thread and being confused by the technical terms might end up telling that your simplistic way of thinking must be the right one, although the whole literature simply proves you wrong.

If you turn any video into interlaced form.... and that's what you keep repeating like it was the new childish challenge against me here, or just a desperate tentative to have me contradict myself so that then you could start acting even more childish and tell everyone that you won "the fight", which doesn't exist anywhere other than in your own mind. I repeat, grow up !

Interlacing is interlacing. Period. If you render even(odd) fields first and the others after you end up with a corrupted optical flow axis. And don't repeat the 1080p24sf stuff again to prove you right because that's not interlacing, although if you studied some basics you would know why motion compensation is needed anyway for scan rate conversion even when you want to achieve a raw multiple of a given source...

You will never ever be able to reconstruct the original progressive signal as it was before and splitting it into fields being displayed in two different groups is interlacing. Telecine pulldown can't do that. What you think or believe just doesn't matter, your statements are wrong. It's not my opinion, anyone studying on the subject should be able to see that.

[darinp2]

A lot of a DVD costs comes from production of the actual discs and boxes. When you cut that out movies become cheaper. People will then buy more.

One of the potential problems I see for the studios with the download model is people only buying movies they are going to watch. I know it may seem crazy, but lots of people buy DVDs and then never get around to watching them, so they are part of the revenue studios have come to expect. The studios got their money if even the DVD never got opened. With download models I think they will pretty much only be able to charge for things that people watch, with few cases of people buying something and then not watching it (or at least not starting to watch it). Like the XBOX360 thing where you have to watch it within a certain amount of time. Most people won't download something they won't get to within that time period, where with DVDs they will buy them in the store knowing that they might not watch it for a month or more (and then sometimes never get around to it).

On DIRECTV I notice that I don't pay for PPV movies until I go to watch them.

This can be an advantage for consumers, but a disadvantage for studios.

--Darin

[Dino]

1080p24sf it's a marketing way of describing just another technique of multipling the video scan rate to reduce flickering, it has nothing to do with interlacing per se.

At the risk of getting in the middle of this argument which has high jacked a
thread about Blu Ray, HD DVD, Microsoft, et al, I would just like to say that
your concept of 24psf is not anything I've heard of.

24psf has been used very successfully by Sony on their F900 series HD CAM camcorders, the first 1080/24p HD camcorder on the market, circa 1999 and still in use today by many, many features, episodics, documentaries and commercials. You've probably heard of Star Wars
Episode II and the new Battlestar Galactica series.

1080/24psf is a way to segment a progressive frame into to halves, not odd
and even lines, not interlacing, as you correctly stated. It allows the tape
format to record an HD progressive signal without even more compression. It also allows this frame rate to stay progressive at 1080X1440 via SDI distribution vs. HD SDI, which was a big cost saver for facilities that already had SDI distribution in place for Digital Betacam.

This is not a consumer format or distributed or transmitted for final release, it is a high end professional intermediate format which, after post-production can be used for film out or transferred to other tape formats and frame rates.

It is not a frame doubling method or marketing trick. Sony also uses 30psf
for standard def. cameras such as the newest DVCAM and Digital Betacam
XD CAM and XD CAM HD models.

Oh yeah, and it flickers lke crazy on most displays!

Dino

[wingzero]

At the risk of getting in the middle of this argument which has high jacked a
thread about Blu Ray, HD DVD, Microsoft, et al, I would just like to say that
your concept of 24psf is not anything I've heard of.

24psf has been used very successfully by Sony on their F900 series HD CAM camcorders, the first 1080/24p HD camcorder on the market, circa 1999 and still in use today by many, many features, episodics, documentaries and commercials. You've probably heard of Star Wars
Episode II and the new Battlestar Galactica series.

1080/24psf is a way to segment a progressive frame into to halves, not odd
and even lines, not interlacing, as you correctly stated. It allows the tape
format to record an HD progressive signal without even more compression. It also allows this frame rate to stay progressive at 1080X1440 via SDI distribution vs. HD SDI, which was a big cost saver for facilities that already had SDI distribution in place for Digital Betacam.

This is not a consumer format or distributed or transmitted for final release, it is a high end professional intermediate format which, after post-production can be used for film out or transferred to other tape formats and frame rates.

It is not a frame doubling method or marketing trick. Sony also uses 30psf
for standard def. cameras such as the newest DVCAM and Digital Betacam
XD CAM and XD CAM HD models.

Oh yeah, and it flickers lke crazy on most displays!

Dino

This Panasonic article from 2002 gives an explanation of how multiplying the scan rate of a progressive film works as well as a rough explanation of Sony Progressive Segmented Frame mode:

ftp://ftp.panasonic.com/pub/Panasoni...ressive-WP.pdf

However it doesn't take into account motion compensation techniques at all, they talk about pulldown only but that's a marketing document made to sell the hardware to professionals in the field, it's a different type of document than one for developers. Also, they say that pulldown "works fine" and not that is a perfect technique nor the only one even when dealing with progressive sources.

This document of an actual deinterlacer chipset with some schematics and bit fields values shows that motion compensation can and is used even with Progressive Segmented Frame videos in a similar way to how it's applied to interlaced video for de-interlacing, although the PSF format is progressive when converting from PSF to the normal full progressive frame format motion compensation techniques are helpful and are being used:

------
http://www.gennum.com/video/pdf/18283DOC.pdf

3.9.1 De-Interlacing Mode (MODE=000)
When set to operate as a de-interlacer the GF9330 can operate as a “stand-alone”
device performing motion adaptive processing. To enable multi-directional edge
and vertical motion detection the GF9330 must be connected to the GF9331 as
described in 3.3 Seamless Interface to the GF9331 Motion Co-processor for
Directional Filter Control.
Segmented frame to progressive frame conversion is also supported in this mode.
This function is performed when the progressive segmented frame input video
format is selected on either the external pins or host interface register STD[4:0].

------


This articles states that PSF mode is used to reduce flickering and is a way to obtain multiples of the source frame rate:

----
http://hd24.com/spectsoft_article.htm

2.Frame type - The letter "p" in the middle indicates the frame type.
In this case it indicates a progressive frame
The standard frame types are p - progressive, i - interlaced, and psf - progressive segmented frame
Notes:
1.A "psf" frame is simply just a progressive frame that just happens to come down the wire as two interlaced fields. This is done to get frame rates such as 24fps above the human flicker perception frequency (i.e. the actual refresh rate becomes 48Hz instead of 24Hz).
2. When an interlaced type is given the following frame rate number actually describes the field rate (or double the frame rate).

----


Another example of using motion compensation for dealing with PSF format in a similar way to interlaced signals can be found on the description of this high end professional product:

-----

http://www.barco.com/projection_syst.../BR_DX-700.pdf
Next-generation LED Processor
DX-700
The DX-700 is a multi-window video processor designed as a versatile front-end to all Barco LED products

All inputs except DVI provide a minimum 10-bit color depth, in either
4:4:4 or 4:2:2 format. An advanced motion-adaptive de-interlacer with
diagonal filter converts interlaced or progressive segmented frame
(PSF) inputs to progressive format.

-----

[xAgonyxScenex]

One of the potential problems I see for the studios with the download model is people only buying movies they are going to watch. I know it may seem crazy, but lots of people buy DVDs and then never get around to watching them, so they are part of the revenue studios have come to expect. The studios got their money if even the DVD never got opened. With download models I think they will pretty much only be able to charge for things that people watch, with few cases of people buying something and then not watching it (or at least not starting to watch it). Like the XBOX360 thing where you have to watch it within a certain amount of time. Most people won't download something they won't get to within that time period, where with DVDs they will buy them in the store knowing that they might not watch it for a month or more (and then sometimes never get around to it).

On DIRECTV I notice that I don't pay for PPV movies until I go to watch them.

This can be an advantage for consumers, but a disadvantage for studios.

--Darin

How is that a problem? Either you forgot to wrap up your thought or I'm missing something. People are going to be more likely to download when they don't have to drive to a store to pick something up, and when its cheaper, because it no doubt will be. The perception is that having a physical product is where most of the cost comes from, and prices reflect that.

Are you suggesting studios are going to be less likely to go strictly digital downloads because of a cash revenue of people who buy and dont watch? Thats kind of silly, the same could be said for someone downloading it and not watching it, but that wont happen until internet speeds are faster.

In terms of XBOX Live, what your talking about is strictly for movies. Movies on XBL is essentially a rental service where you download 6-7gig 720p HD movies. It's different for everything else. once you download it, it's yours.